consiletter 7

Thermal hazards: identify, assess, control…

Part III: Help! I have a runaway reaction!

In our last two consiLetters, we showed how to determine safety-relevant parameters for safe thermal handling of substances. In combination with additional data about the actual reaction, such as reaction enthalpy or heat flux, whose measurement we will cover in a later consiLetter, measures for normal operation can now be developed based on the TRAS 410 technical rules for plant safety, which ensure safe operation of the plant.

adiabatic reaction calorimeter

But what about when the operation deviates from normal operation, such as due to a loss of cooling or incorrect dosing, and an unwanted rise in temperature results that could possibly initiate additional thermal process (e. g. decomposition reactions). In these cases it is important to assess the possible impacts and safely control them. First, realistic malfunction scenarios (worst case) scenarios should be defined. These deviations from normal operation should then be examined in an adiabatic reaction calorimeter (e.g. VSP2) in order to assess possible impacts. The adiabatic conditions in the laboratory test are essential for scale-up, since a laboratory reactor generally has a much larger surface-to-volume ratio than a production reactor. The relative heat loss in a non-adiabatic laboratory reactor would therefore be too great and would lead to a non-conservative result.

With the adiabatic VSP reaction calorimeter (Fig. 1) reactions that deviate from normal operation (e.g. higher heating temperature, cooling failure, improper charging, etc.) can be tested under safe conditions at a 100 ml scale. In the process, reactants are placed in what are usually a closed, very thin-walled test cell with a volume of 120 ml, and this well-insulated cell is installed in a pressure vessel with a permissible operating pressure of 120 bar. The reaction mixture is then warmed with the test cell heater to the desired starting temperature, from which the further increase in temperature results solely from the energy released by the reaction. To minimize the heat exchange with the environment, the temperature in the pressure vessel is continuously held at the temperature in the test cell with the use of a heating system, which creates a “quasi” adiabatic system that very closely replicates the conditions in a real boiler.

In a VSP experiment, not only can the maximum (adiabatic) temperature and pressure rise of a reaction be specified for the tested deviation from normal operation (Fig. 2) but also the respective speeds of temperature and pressure rise independent of the temperature. Additionally, a prediction regarding the formation of permanent gases can be made via the cooling curves. Based on this data, for one thing, a suitable countermeasure can now be developed to minimize the impacts of the deviation (e. g. emergency cooling), and for another, the data can be used to size the pressure relief equipment (safety valve or rupture disk) so that the plant is protected from an impermissible pressure rise in the case of a runaway reaction.

If we can assist you with a similar issue, please contact us. Our experts will be happy to advise you in specifying the test conditions and can also support you early on at the safety briefing stage.

Your contacts

consiletter 6

Thermal hazards: identify, assess, control…

Part II: Help! My process temperature exceeds the limit temperature

In our last consiLetter we showed how you can obtain a limit temperature for safe handling and use at an early stage in the development of a product or process and using a very small amount of substance in a screening test in the differential scanning calorimetry (DSC) cell. In some cases, however, the limit temperature specified with the DSC is very low and significantly below the desired process temperature. Because changing the process is often extremely difficult, we recommend in these cases more in-depth analysis of thermal stability.

adiabatic heat-pressure accumulation test

One of these additional assessments is the adiabatic heat-pressure accumulation test. Here, approximately 100 g of the substance is weighed out in a double-walled vacuum jacket (Dewar) with a reflective coating. This Dewar vessel is then inserted into a stainless steel autoclave, which is placed for the duration of the test into a massive aluminum block furnace, Fig. 1. This setup ensures that there is no exchange of heat or substances with the surrounding environment (“adiabatic conditions”). To replicate the actual process as closely as possible, stirring or dosing of additional components is possible. In addition to the temperature in the sample, the pressure in the headspace of the autoclave is also measured.

adiabatic temperature rise (∆Tad)

An adiabatic measurement of this type not only simulates the worst-case scenario of a malfunction, or failure of cooling, but the adiabatic temperature rise (∆Tad) can also be derived from the test. The adiabatic temperature rise indicates the temperature difference by which the sample may heat, independent of the volume or mass of the sample. Another parameter of interest for a thermally safe process, which is likewise obtained from the measured data, is the adiabatic induction time (TMR_ad). This shows the time span within which, under adiabatic conditions, the maximum rate of temperature rise is reached.

The adiabatic decomposition temperature for 24 hours (“Adiabatische ZersetzungsTemperatur” AZT24) describes the temperature at which the process requires 24 hours, under adiabatic conditions, to reach the maximum rate of temperature rise. In accordance with TRAS 410, the limit temperature for safe handling (T_exo) was selected at 10 K below the AZT24.

Figure 2 shows as an example of the temperature profile (blue), as well as the temperature rise rate (rot) from a test with a peroxide. First, the adjustment of temperature to the set oven temperature is observed. Afterward, an initially subtle temperature rise is observed. Due to the adiabatic test process, the heat produced cannot be dissipated and the product continues to become warmer, ultimately transitioning to an exponential temperature rise. The maximum temperature rise rate is detected at a temperature of 197 °C. The AZT24 is obtained by going back 24 hours from this point. For the decomposition of the peroxide, this is 90 °C. The Texo is therefore arrived at by subtracting the safety margin of 10 K, for 80 °C. Consequently, the peroxide can be safely handled at a temperature of up to 80 °C.

The observed adiabatic decomposition of the peroxide not only leads to an adiabatic temperature rise of about 210 °C but also is accompanied by pressure buildup. In the next issue of the consiLetter we will show you what you can do when the pressure buildup identified in an adiabatic test exceeds the design limits of the plant.

If we can assist you with a similar issue, please contact us. Our experts will be happy to help you.

Your contacts

consiletter 5

Thermal hazards: identify, assess, control…

Part I: Quick screening with very small amounts – along with fast help for R&D

Many processes in the chemical industry lead to the release of energy in the form of heat, which in an uncontrolled release may lead to a self-accelerating reaction. The heat that often accumulates here is a possible catalyst for additional energy-rich secondary and consecutive reactions or for the simple decomposition of the substance. This so-called thermal “runaway” often cannot be controlled through (counter-) cooling. The occurring decomposition products may ultimately lead to an uncontrolled pressure buildup with far-reaching consequences.

Differential scanning calorimetry (DSC)

Differential scanning calorimetry (DSC) provides an in initial indication of the hazard potential of substances and compounds. Here, a small sample of approx. 10 mg is heated in comparison with an inert reference. Endothermal or exothermal processes are measured here as differences between the thermal release of the sample and the reference, plotted against the temperature (“Thermogram”), and enable conclusions to be drawn regarding the thermal potential of the substance. The start of the effect, known as the “onset temperature,” can be inferred from the thermogram. Figure 1 shows an example of a thermogram. On the basis of the identified“onset temperature,” a limit temperature “Texo” can be specified up to which a substance or mixture of substances is generally safe to handle. In accordance with the technical regulation for plant safety TRAS410, this limit temperature is set 100 K below the “onset temperature” of the first exothermic setting.

Using the thermogram, the energy content can furthermore be determined through integration of the signals. This is an important parameter for the REACH Regulation, the European Agreement concerning the International Carriage of Dangerous Goods by Road (ADR) transport law and the assessment of chemical processes (TRAS410). In these regulations, limits are stated for the release of energy which, when exceeded, assume certain thermal potentials for a specific substance or mixture of substances. For instance, no explosive properties are assumed for a substance with an energy release of less than 500 J/g up to 500 °C. In our experience, an assessment of the thermal hazard potential for most of the samples tested can be conducted using these simple and quick methods – and using a minimum sample size. It is therefore the method of choice, especially at an early stage of chemical development and laboratory synthesis.

DSC for sugar

Figure 2 shows as an example the DSC for conventional sugar. First, endothermic caramelization is observed (blue segment). Starting at a temperature of about 220 °C, the exothermic decomposition can be seen (red segment). The T_exo is thus up to 120 °C. Consequently, sugar can be handled at a temperature of up to 120 °C without a thermal hazard. As an additional safety-related classification, sugar, with an energy release of about 450 J/g, is not viewed as an explosive substance. As this simple example shows, with about 10 mg of a substance, the significant parameters for thermal stability and additional information such as melting and boiling points can be obtained.

In the next issue we will suggest things you can do if the limit temperature Texo identified in the DSC could be exceeded in a chemical process. If we can assist you with a similar issue, please contact us. Our experts will be happy to help you.

Your contacts

consiletter 4

Identifying very low vapor pressure (e. g. of solids)

No data, no market: That’s what it says in Article 5 of the General obligation to register and inform requirements for chemicals covered by the new REACH Regulation, which are currently being introduced incrementally. Because of this, determination of vapor pressure will soon play a larger role in the chemical industry. Based on REACH, determining vapor pressure falls under the standard requirements for physiochemical studies (see REACH Regulation Annex VII 7.5) and is therefore required for each registration of a substance.

effusion method

Vapor pressure can be determined using the various methods described in the OECD Guideline 104. With the effusion method/vapor pressure balance, vapor pressure can be determined from solid and liquid substances in a pressure range from 10^-3 to 1 Pa (or 10^-5 – 10^-2 mbar). This range is especially important for medium- to low-volatile substances.

The principle of the effusion method is based on a molecular beam that is emitted through the defined opening of a Knudsen effusion cell and that condenses on a target plate cooled with liquid nitrogen. The vapor pressure of the test substance can be computed with the Knudsen effusion equation from the increase in weight of the target plate, which is measured using a highly sensitive microbalance (see Figure 1). In Figure 2, the vapor pressures of benzoic acid measured at various temperatures are plotted over the inverse temperature. The Antoine parameters can be determined from the straight lines formed in the semi-logarithmic diagram, with which the vapor pressure at other temperatures can be extrapolated. The unique aspect here is that substances with a very low vapor pressure (e. g. solids) can be reliably measured this way.

In addition to the described vapor pressure balance, the consilab laboratory features a second device that determines vapor pressure from the loss of mass caused by the effusion (recommended measuring range according to OECD 104: 10^-10 to 1 Pa).

As a certified GLP testing facility, consilab naturally also offers determination of vapor pressure in the form of a GLP study. In addition, our laboratory is equipped with all other physiochemical methods required for the new approval of substances.

If we can assist you with an issue related to determining vapor pressure, please contact us. Our experts will be happy to help you.

Your contacts

consiletter 3

Designing Safety Valves for the fire case?

Frequently, the question arises whether safety valves on containers should also be designed for the case of fire. In TRB 403 of October 1998 „Equipment of Pressure Vessels – Devices for recognizing and delimitating pressure and temperature“ the response to this question still was a definite NO. Here, paragraph 3.2.3 says: „When designing safety facilities, it is NOT necessary to consider the fire case.“ In the subsequent guidelines, AD 2000 Information Sheet A 403 with the same title, paragraph 3.2.3 the June edition 2001, however, contains the following statement: „Taking the external fire case into consideration MAY be necessary when designing safety devices against overpressure.“ Meaning: Today immediate answers to this question can no longer be given with absolute certainty.

In API 521 or ISO 23251, the process of calculating heat input for the fire case for different containers under varying surrounding influences is described in detail. Here, it is not only deemed important whether the container is equipped with a heating jacket but also, how quickly fire fighters can be on the site, or, whether there is a sprinkler system. Figure 1 shows the heat input depending on the filling level of a 10 m³ container filled with acetone being heated with differing steam temperatures. For the sake of comparison, the heat input due to fire as calculated by normative formulas is also displayed.

In the most sites of chemical plants, as is also the case in Industrial Park Hoechst, fire fighters arrive on scene in less than ten minutes. Thus, cases in which the worst case scenario need be assumed are indeed rare. The fact that, by standard, many plants possess sprinkler systems, significantly even more reduces heat input due to fire in most cases. As the diagram shows, the design case of „External Heating “ at 6 bar steam normally yields much more heat input than the fire case.

This means, that often, standard containers with heating jacket need not be designed for the fire case if external heating by steam has already been considered. However, such consideration must be given in the case that the heating jacket is designed to function as isolation in the event of underfiring. Safety briefings for containers without heating jackets thus should take special care to ensure that the case of underfiring is effectively excluded, for example, by designing drain channels to prevent accumulation of inflammable/combustible liquids underneath the container.

If you have any questions regarding similar cases, please do not hesitate to contact us. Our experts are always very happy to assist you.

Your contacts

consiletter 2

Determining the Self-Ignition Temperature of Dust Deposits

Most solids produced or handled in the chemical industry are powders. These substances have a large surface/volume ratio enabling oxidation processes to occur already at room temperature. Due to the poor heat dissipation of bulk materials, the substance might heat up or even ignite on its own. According to the relevant standards and regulations “self-ignition” is defined as a temperature increase by more than 60 K (compared to the initial storage temperature). If the temperature increase is less, this is denominated “self-heating”.

Generally, the self-ignition temperature is determined by several isoperibolic heat storage tests at different constant temperatures and different sample sizes. This method takes a lot of time and needs large amounts of the substance. However, our customers frequently can only provide a relative small amount of the material for the required investigations.

To gain an economy of time, costs and material for our customers, we offer the advantages of adiabatic wire basket tests. For this purpose the sample is filled into a wire basket of a volume of 400 ml for isoperibolic storage over 24 h. If no self-heating occurs within this period of time, the oven temperature is increased so slowly, that a uniform temperature distribution in the sample volume can be assumed. As soon as the sample temperature exceeds the oven temperature, the oven temperature is regulated to trace the sample temperature. Thus, adiabatic conditions are provided in this phase of the test.

Figure 1 shows the experimental set up as used in our laboratory. Additionally to the core temperature, the temperature at the surface, and the temperature midway from center to surface are measured.

A typical kinetically evaluated test is depicted in Figure 2. The sample was tempered at 50°C and subsequently stored at 100°C for 24 h. As no self-heating occurred, the oven temperature was increased by 1 K/h. At approximately 152°C the sample temperature exceeded the oven temperature and self ignition came to pass.

This kind of experimental procedure induces advantages for our customers. Due to the adiabatic conditions the values for the activation energy EA, the Arrhenius prefactor k0 (from the linearized slope of the temperature rise curve caused by self-ignition), and the thermal conductivity a can be derived from the tests. From these values plus the geometry of the sample, the required self-ignition temperature can be evaluated iteratively using the “theory of stationary heat explosion” by D. A. Frank-Kamenetskii. In contrast to isoperibolic experiments, the self-ignition temperature for larger bulk sizes can be directly derived from the measured values with respect to the geometry. An extrapolation procedure is not required, and the activation energy is considered. The obtained temperatures are significantly more reliable, which leads to a considerably higher safety level for our customers.

If we can support you at a similar task, please feel free to contact us. Our experts will be happy to place our services at your request.

Your contacts

consiletter 1

Safeguarding of a supply line for pressurized air

As in many of our customers’ laboratories, the apparatuses in the laboratories of consilab have to be supplied with technical gases, too. In most of the cases, the tasks of pressure vessel safety are similar and the questions recur. For this reason, the assembly in the consilab laboratory shall be described as an example.

Commonly, compressed gas cylinders can only be operated when equipped with a pressure reducing device. But what is going to happen, if the pressure reducing device of a new, completely filled gas cylinder is damaged or fully opened when connected to the laboratory equipment? The total pressure of the cylinder could propagate via open pipelines into the apparatuses and eventually destroy those. To prevent this, a safety valve is to be installed into the feed line, to set the maximum pressure of the system to the maximum allowable pressure of the connected apparatuses.

The sizing of safety valves is part of the range of services offered by consilab. First of all, the mass flow rate to be discharged has to be calculated. This is the mass flow rate of gas which can flow from the gas cylinders into the connected apparatuses. Basis of the calculations are the pressure difference between maximum pressure in the gas cylinder and the set pressure of the safety valve or the maximum allowable working pressure of the apparatus. The geometric assembly of the feed pipe (diameter, bends, length of pipe from pressure source to the place where the safety valve is connected) can be taken into account for the calculation. In many cases the seat area of the pressure regulator, which could open completely due to a full break of the membrane, is most relevant for the result.

In the next step a safety valve has to be chosen. The dischargeable mass flow at set pressure has to be equal or higher than the mass flow to be discharged which was calculated in the first step. On the other hand, the safety valve should be as small as possible, since the gas feed lines are usually very small and there is not much space in a fume hood. But not every safety valve is applicable to reliably venting the required gas mass flow.

Because of the high pressure difference of the compressed gas cylinder (300 bar at consilab) and the apparatus (30 bar at consilab) and the relatively low pressure difference between set pressure and ambient pressure, the required venting area is accordingly high.

In the actual case a safety valve by Leser, size ½”/ ½”, type 437, is sufficient to protect the feed line in size DN6. The easiest way to connect the safety valve to the feed line seems to be to insert a branch tee and expand the line behind it to the size of the safety valve inlet. However this is not permissible. Standards like ISO 4126 regulate that there must not be a diameter smaller than the safety valve inlet diameter in the line between the room to be protected and the safety valve itself. Additionally, the pressure losses in the line to the safety valve are to be restricted to 3% of the set pressure. To meet these criteria, the diameter of the feed line has to be expanded to the safety valves inlet size before the branch tee. An example of an ideal assembly for this task is shown in the photo.

If we can support you at a similar task, please feel free to contact us. Our experts will be happy to place our services at your request.

Your contacts